Move Heroku Postgres Without Losing Salesforce Sync

Why does Salesforce sync break when Heroku Postgres moves?

Heroku Connect is a Heroku add-on. It reads and writes Salesforce objects into a schema inside a Heroku Postgres database, usually named salesforce, and tracks changes with its own trigger and log tables. That design works while everything lives on Heroku. It does not travel.

If you dump and restore the database somewhere else, you get a copy of the synced tables but not the process that keeps them current. Records written in Salesforce after the dump never arrive. Rows your app writes to the copied tables never reach Salesforce. The data looks fine on day one and drifts from there.

What should you inventory before the move?

Before touching the database, write down what Heroku Connect actually does for you today. Most teams find the list is longer than they remembered.

• Every mapped Salesforce object and the fields on each mapping.
• Which mappings are read-only and which are read-write.
• Tables, columns and indexes your app queries directly in the Heroku Connect schema.
• Any app logic that depends on Heroku Connect system columns, such as the record state or external ID columns.
• Jobs, reports or other services that read the same tables.
• How often the business expects changes to show up on each side.

This inventory becomes the spec for the replacement sync. If the new tables keep the same shape, application changes stay small.

Who handles which part: Build.io or Stacksync?

The work splits cleanly along the line between platform and data sync.

Because the two pieces are independent, each team can test its part without waiting on the other, and the joint plan only has to agree on the cutover date.

What are the steps for a clean cutover?

1. 01
   Map the current Heroku Connect schema

   Export the list of mapped objects, fields and directions. Note every table and column the app reads.
2. 02
   Provision the new Postgres database

   Build.io sets up the target platform and database. Keep table and column names compatible with the Heroku Connect schema.
3. 03
   Connect Stacksync to Salesforce and the new database

   Use the <a href="/connectors/salesforce">Salesforce connector</a> and a Postgres connection. Configure each object as one-way or two-way to match what Heroku Connect did.
4. 04
   Run the initial load and let it catch up

   Stacksync backfills the Salesforce records into the new tables. Leave Heroku Connect running on the old database so production is unaffected.
5. 05
   Validate counts and sample records

   Compare record counts per object and spot-check field values. Test a write from Postgres to Salesforce and from Salesforce to Postgres.
6. 06
   Switch application traffic

   Point the app at the new database through Build.io. Watch sync activity for errors in the first hours.
7. 07
   Retire Heroku Connect

   Once the new setup is stable, turn off the Heroku Connect mappings and remove the add-on.

What changes after Heroku Connect is gone?

The day-to-day work of the app stays the same: it reads and writes Postgres tables. What changes is the sync engine behind those tables. Stacksync reads changes from Salesforce and the database and applies them to the other side, so engineers stop maintaining trigger logs or polling jobs.

Teams also gain the option to sync more than Salesforce. If sales runs on HubSpot or finance runs on NetSuite, those systems can sync into the same Postgres database. For a full comparison with Heroku Connect, see our Heroku Connect alternative page.

If you are planning a Heroku exit and want to walk through your current mappings, book a demo and bring your list of objects.